Whoops: Google’s ad for new ChatGPT rival Bard shows the AI chatbot giving an inaccurate answer
- Google recently revealed its experimental AI chatbot, Bard, a rival to OpenAI’s viral ChatGPT.
- In an ad for Bard, the chatbot gives an inaccurate answer to a question about the James Webb Space Telescope.
- Bard is the latest player in the Big Tech battle for consumer-facing AI.
Thanks for signing up!
Access your favorite topics in a personalized feed while you’re on the go.
Google’s ad for its new experimental AI chatbot, Bard, proves that artificial intelligence may not be taking over anytime soon.
The ad depicts the chatbot giving an inaccurate answer to a question about the James Webb Space Telescope (JWST), Reuters first reported.
In the ad, a user asks Bard, “What new discoveries from the James Webb Space Telescope can I tell my 9 year old about?”
“JWST took the very first pictures of a planet outside of our own solar system. These distant worlds are called ‘exoplanets.’ Exo means ‘from outside,” Bard answers, among other “facts.”
However, the first photograph of an exoplanet was not taken by the JWST. It was taken in 2004 by the European Southern Observatory’s Very Large Telescope, according to NASA. 2M1207b was the first directly photographed exoplanet, per the space agency.
Bard’s other two answers — one about JWST’s spotting of “green peas” galaxies this year and one about its images of a galaxy that’s over 13 billion years old — are accurate.
In a blog post from earlier this week, Google CEO Sundar Pichai wrote that Bard uses “large language models” and information from the internet “to provide fresh, high-quality responses.”
It’s possible the error could be because the chatbot used information from a NASA blog post about the JWST announcing “the first time” astronomers used the telescope “to take a direct image of a planet outside our solar system.”
“This highlights the importance of a rigorous testing process, something that we’re kicking off this week with our Trusted Tester program,” a Google spokesperson told Insider about the error. “We’ll combine external feedback with our own internal testing to make sure Bard’s responses meet a high bar for quality, safety and groundedness in real-world information.”
Pichai said Bard is being initially released with the company’s “lightweight model version of LaMDA” — which stands for Language Model for Dialogue Applications — Google’s conversational technology that was trained on dialogue. Users can use the chatbot and offer Google feedback, Pichai said.
Bard was released to “trusted testers” this week, and Pichai said the public will have access in the coming weeks.
Since the release of OpenAI’s chatbot ChatGPT in November, tech giants have been racing to create consumer-facing AI. Google issued a “code red” to staff, warning of the potential oncoming threat to its search engine.
Even with the introduction of Bard — and other AI-based technology — it doesn’t seem to be going well for the company. Shares of Alphabet, Google’s parent company, have fallen by over 8% since Google’s AI showcase in Paris this morning.